The skinfold thickness is a much-used measurement for monitoring adiposity in a wide range of medical, health, occupational and sport science disciplines. Misconceptions abound, however, in its use, particularly that of purportedly predicting body 'fat' as opposed to 'adipose tissue'. To obtain data to investigate body composition and the extent to which anthropometry can be justifiably used to predict whole-body adiposity, an extensive dissection study was undertaken on 34 cadavers. In addition, to pre-empt questions on the applicability of cadaver data to living subjects, 40 elderly in vivo subjects of the same age range were compared with the cadaver population. No significant macro-morphological differences were found between males or females in the morbid and in vivo groups. Significant findings affect our previous understanding of the predictability of whole-body 'fat'. Skinfold compressibility was by no means constant; skin thickness varied with location in both sexes, females having thinner skin than males; there were significant sex differences in adipose tissue patterning. An identical thickness of adipose tissue did not necessarily contain the same concentrations of fat. Despite this variability, a relationship was demonstrated between aggregate skinfold measures and subcutaneous adipose tissue mass (as opposed to subcutaneous fat), this relationship being more evident in men. A strong relationship was found between subcutaneous adiposity and whole-body adiposity, and between direct skinfold depth measures and whole-body adiposity. The amount of visceral adipose tissue was the same in men and women, but in the men this represented a greater proportion of their total body adiposity. Further, the use of waist-to-hip girth ratio (WHR) was identified as an important predictor of health risk. These findings demonstrate that it is not sustainable to introduce a non-quantifiable error by transforming anthropometric values (skinfolds) into predictions of percentage body fat. If subcutaneous adiposity can be predicted, then an excellent indication of overall adiposity could be obtained. Currently, skinfold measurement can yield a reasonable indication of comparative subcutaneous adiposity (better in men than in women). In neither gender is this prediction completely reliable due to both inter- and intra-individual differences in the skinfold measurement procedure.